OpenAI does not like to be left out. The week after Anthropic announced Claude Mythos Preview—an AI model that has put governments around the world on edge because of its potential ability to hack into banks, energy grids, and military systems—OpenAI shared a program that is uncannily similar. And just like Anthropic did with its model, OpenAI has, for cybersecurity purposes, restricted access to this new bot, called GPT-5.4-Cyber, to a small group of trusted users.
This sequence has become something of a pattern: First Anthropic will make an announcement, and then OpenAI will follow suit. Last year, Anthropic launched Claude Code, an AI coding tool. A couple of months later, OpenAI came out with its own version, Codex. When Claude Code had a breakout moment in January, OpenAI responded with two major updates to Codex alongside a press blitz for the product. And earlier this month, OpenAI released a version of Codex that allows it to use other apps on your desktop—similar to an existing Anthropic tool called Claude Cowork.
Until recently, Anthropic—founded by a group of former OpenAI employees in 2021—played the role of younger brother. OpenAI kicked off the entire AI boom with the release of ChatGPT, and has had more users, funding, and name recognition ever since. But Anthropic has been riding high on the explosive popularity of Claude Code and booming sales of its AI models to large corporations. The firm’s showdown with the Pentagon has also helped vault it into the public eye. In early April, Anthropic said its revenue rate had hit $30 billion a year—appearing to surpass OpenAI’s.
In its public messaging, OpenAI has been indifferent or even somewhat derogatory toward Anthropic. Last week, when OpenAI released its newest model, GPT-5.5, the announcement was paired with direct and veiled references to how it beat out Anthropic’s latest, Claude Opus 4.7. But internally, the firm is seemingly on edge. In a recent leaked company-wide memo, Denise Dresser, OpenAI’s chief revenue officer, felt the need to address one particular competitor: “Here are a few things worth keeping in mind, especially on Anthropic.” The rival firm’s product offerings are narrow, Dresser wrote, and “their story is built on fear,” referencing Anthropic’s loud messaging about the dangers of AI. “Our positive message will win over time.” (OpenAI, which has a business partnership with The Atlantic, did not respond to a request for comment. Anthropic also did not respond to a request for comment.)
If imitation is the sincerest form of flattery, OpenAI’s actions are especially telling. At every turn, OpenAI has appeared eager to copy the success of its rival. For starters, as Anthropic’s explicit focus on mitigating the risks of AI has apparently won the trust of many consumers, OpenAI has imitated many of its rival’s safety initiatives. In early 2026, after Anthropic published a major update to Claude’s “Constitution,” a document that tells the AI model how to behave, OpenAI launched a major campaign around its equivalent document.
But OpenAI’s most important, Anthropic-esque pivot has been in its business model. Early on, these two companies made fundamentally different bets on how they would eventually make money. OpenAI positioned itself as a consumer behemoth, hoping to capitalize on ChatGPT’s hundreds of millions of users. Last fall, the company launched the AI-video app Sora and an AI-powered web browser. OpenAI has made forays into e-commerce and is testing ads in ChatGPT. Every now and then, the company teases the AI device that it is developing with the former Apple designer Jony Ive. Anthropic, meanwhile, has focused on the less flashy goal of selling its AI tools to businesses and software engineers.
Despite OpenAI’s numerous advantages, Anthropic’s focus on code and business customers seems to be winning. Although OpenAI is worth more based on the most recent fundraising rounds, Anthropic now has a higher valuation—more than $1 trillion—in some private markets. Anthropic’s explosive growth is particularly important as the two companies both race to go public, in turn accessing a huge pool of new investors, and try to prove they will eventually be profitable. (Both companies still have a long way to go in that regard.)
OpenAI is now eager to catch up. In December, OpenAI hired Dresser, a former CEO of Slack, to pursue more business customers. In late January, Altman gathered several major executives for a lavish dinner in San Francisco to preview all of the business offerings his company was planning, according to The Information. The company has since made a blitz of announcements around coding tools and enterprise AI offerings, including a new set of “Frontier Alliances”: partnerships with several of the world’s premier consulting firms, including McKinsey & Company and Boston Consulting Group, to accelerate enterprise adoption of ChatGPT. In mid-March, another internal OpenAI memo reportedly stated that the company needed to eliminate “side quests” and focus on the enterprise and coding markets. Anthropic’s success in those areas, the memo stated, should be a “wake-up call” for OpenAI. The firm also scrapped Sora and has been aggressively advertising and messaging about Codex for months now. “I am happy everyone is switching to Codex,” Altman wrote on X earlier this month.
OpenAI’s pivot to its enterprise business has not been total. It did, for instance, recently shell out reportedly hundreds of millions of dollars to acquire a niche tech podcast. And Anthropic, for its part, has had to take some cues from OpenAI—notably by making big and expensive data-center deals, such as an expansion in its partnership with Amazon Web Services. Anthropic’s CEO, Dario Amodei, has previously insinuated that OpenAI has made such deals “because it sounds cool.”
Which company will win the AI race is anybody’s guess. Regardless, OpenAI’s embrace of the Anthropic business model makes one thing abundantly clear: For all the wonder and change that generative AI brings as a technology, there hasn’t been any real innovation in the business models of Silicon Valley. For decades, most tech companies have succeeded by either selling ads (the route of Meta and Google) or selling enterprise tools (like Salesforce and Slack). One day OpenAI or Anthropic might cure cancer and remake the world, but for now they still have to pay the bills.
